Changing It Up

Changing It Up


10 min read

Dan Hymowitz Former Senior Advisor

Posted on: 27th June 2016


The meeting was getting tense. In December 2014 I was sitting with the head of Monrovia’s Ebola response task force, Sonpon Sieh, as experts from an international agency explained how Monrovia’s Ebola coordination system should be set up. There was already a system in place, but they said Sieh should scrap that system and build a new version – a new organisational chart, new reporting lines with new types of meetings – based on best practices from other crises they’d worked on. After a long debate in which Sieh defended the current system and the experts advocated a new version, Sieh chose not to take their advice.

The irony is that only a few weeks later, Sieh, working with international partners, did indeed revamp the Monrovia Ebola response system,1 but in a completely different way from how those advisers had suggested. This change, which helped Liberia end its Ebola epidemic, was tailored to the particular challenges of Ebola in Monrovia at that moment. The experts, in contrast, had recommended a solution that had been effective in rural Ebola outbreaks in other countries, but was unlikely to help Liberia deal with the particular challenges of managing Ebola in an urban setting.

There’s a lot of discussion in global development today about the need to work in more context-specific, adaptive ways rather than importing best practices. Some call it ‘Problem-Driven Iterative Adaptation (PDIA)’,2 some ‘politically smart, locally led development’3 and some ‘Doing Development Differently’.

There’s a risk those of us working on delivery – governments and international partners alike – agree with this principle, but fail to apply it in practice. As in other areas of development, there’s comfort in sticking to what we know, so we tend to fixate on delivery mechanism ‘best practices’.

In delivery this typically means putting a team of people in the centre of government who work from a manual, using tracking and reporting templates copied from other countries. But in our experience, this cut-and-paste approach doesn’t work. It takes locally customised mechanisms to drive the implementation of reforms successfully.

We’ve found that the only way to get to a tailored version that works in a given context, is to start with key principles, and then test different approaches to adapt them locally.

This is the third article in our ‘Art of delivery’ series about common mistakes and emerging lessons in delivery work today. In our first piece we argued that too often delivery mechanisms fail to improve policy implementation because there’s too much focus on technical processes such as performance monitoring, and not enough on the artistic facets of delivery such as using political authority and incentives – though of course the answer, as one person pointed out about that paper, is that you need both. In the second we focused on the challenge of prioritisation in a political context.

This third paper builds on the first by looking in more detail at the experimental and iterative way in which we have adapted delivery mechanisms in order to find a solution that takes into account those more artistic elements of implementation.

In our experience you are unlikely to create an effective mechanism at the first attempt; you’ll need to learn and adapt. Sometimes adapting means a modest change – a ‘minimally invasive procedure’ such as changing how a meeting is run. Or you may need to make bigger changes when the system no longer works or the problems shift – ‘major surgery’ such as scrapping a mechanism altogether and setting up something new. This means accepting that, in order to build effective delivery systems, you might need to change not only the plan itself but also the delivery process mid-course.

The rest of this paper gives examples of adaptations to delivery mechanisms we’ve worked on in Africa, from minimally invasive change to major overhaul. This list is not exhaustive, nor does it offer a model for governments and international partners on how to undertake an adaptive approach in practice. But to inform the discussion on PDIA, we hope it provides useful examples of the types of adaptation and mid-course shifts in approach that practitioners should be exploring.

Building Roads in Liberia

The original mechanism

Liberia’s Ministry of Public Works’ Project Coordination Unit (PCU) is a ministry-level delivery unit. Initially the PCU was set up to monitor the progress of priority projects; in the words of Public Works Minister Gyude Moore, ‘to bring the same accountability within the Ministry that the President’s Delivery Unit had brought across sectors’.

Why a change was made

Early on, the PCU was struggling to fix problems in project implementation. Project engineers weren’t cooperating with the unit, which they saw as a watchdog whose role was to criticise their work.

The adaptation and impact

An adjustment to a weekly Friday meeting between the PCU and the Minister helped make the system more effective at solving problems. Initially this meeting had served as a progress report and information- sharing session for the Minister.

The meeting evolved into a proactive problem-solving session including not only the Minister, but also other key players such as the Deputy and Assistant Ministers. The PCU would bring ground-level information – often a combination of data and photos of projects – to the group and together they would figure out how to troubleshoot particular problems.

Recent progress on Liberia’s road from Fishtown to Harper illustrates the power of this mechanism. Earlier this year, a PCU team visited the road – roughly a two-day drive from the capital and found the project was behind schedule. Issues included that the contractor didn’t have enough equipment and the laboratories weren’t functional.

The PCU brought a detailed report with photos to the weekly meeting with the Minister. Shortly thereafter, the Minister travelled to the project with representatives from the project’s main funder, the African Development Bank, and issued an ultimatum: finish at least 10 kilometers of road formation before the end of March 2016, or the contractor would not be allowed to bid for the second phase of the project.

The turnaround since Minister Moore’s visit has been striking.The contractor finished more than 25 kilometres by the end of March, in part because it brought in more equipment. And the faster pace has continued since.

Transformational Initiatives in Rwanda

The original mechanism

Imihigo is a practice in which government institutions and officials in Rwanda sign up to contracts with their superiors. A modified version of a Rwandan cultural tradition, Ministers sign these contracts and submit them to the President committing them to delivery targets. In recent years, Imihigo has contributed to improvements in health such as a big increase in health-insurance coverage, road construction and agricultural production.

Why a change was made

A few issues have limited Imihigo’s impact on development outcomes. First, budgets did not always match the targets – too often projects under Imihigo haven’t received the funding they needed. Second, the contracts have mainly included outputs rather than transformational initiatives (which require more coordination across institutions). For example, since 2012, Rwanda’s Ministry of Agriculture had planned the Gishari flower park, a priority project under Imihigo aiming to boost the country’s exports. But the Ministry of Infrastructure did not prioritise and finance the road needed to transport the flowers to the airport. So while the Ministry of Agriculture had successfully delivered its Imihigo commitment, the country didn’t receive the full benefit of the initiative.

The adaptation and impact

To tackle these problems, in 2015 the government shifted oversight of Imihigo from the Prime Minister’s Office to the Ministry of Finance. The Finance Ministry has since introduced the ‘Joint Imihigo’, which is focused on ‘transformational’ outputs in seven cross-cutting priorities such as exports and job creation. The idea is that key elements of multifaceted initiatives, such as the road to the flower park, will no longer be missed. And with the Finance Ministry overseeing the process, financing gaps, which have tended to be a constraint, can be dealt with faster.


National and Local Accountability Together

We’ve also seen that experimenting with additional delivery mechanisms can sometimes improve results.

The original mechanism

In Sierra Leone prior to Ebola, the President received regular updates on the implementation of priority projects through a delivery mechanism that was based in the President’s Office, working closely with priority sector ministries and agencies.

Why a change was made

Sierra Leone has expanded its delivery architecture to the local level, building on the lessons from its response to Ebola. The government found that the Ebola response was most effective when local coordination bodies, the District Ebola Response Centres, had the autonomy to track progress and resolve issues at a district level.

The national response structure relied on the districts to track their own data, identify issues and develop locally-owned solutions. This delivered better results, more quickly. As the government thought about its recovery plans after Ebola, they recognized the value of applying the same concept to driving better outcomes in sectors like education, health and agriculture.

The adaptation and impact

Sierra Leone incorporated district-level Delivery Teams into a new delivery mechanism designed to drive the President’s Post-Ebola Recovery Priorities. At the centre of government, a Delivery Team holds weekly ‘Delivery Forums’ with the President’s Chief of Staff to review progress with senior government officials from the responsible ministries. District Delivery Teams also host their own ‘District Delivery Forums’, chaired by District Council Chairmen, with officials from ministries that work in the district.

The district structure is both parallel to and integrated with the central system. Parallel in that it offers a direct accountability mechanism to track and follow up on activities. Integrated in that the data feeds upwards to the centre, thus providing another information source. This can be helpful. For example, one district was able to report that two international partners were providing the same water project to the same hospital – a duplication of much needed resources. The Ministry hadn’t yet identified the duplication and the report through the delivery system allowed for an expedited solution.

These are still early days for the district delivery forums, but results have been promising, and the government is now exploring an even more local process. It is working with local chiefs who will be responsible for delivery targets in their communities.

Solving a Political Challenge at the Highest Level

The original mechanism

Additions to delivery architecture can also prove effective, as shown in one country where a delivery unit was managing the President’s delivery agenda. In the first couple of years, the unit helped deliver progress in several priority sectors.

Why a change was made

After a while, the system wasn’t working as well. Tensions between the Minister of Finance and the Chief of Staff – both of whom were rising stars in the governing party and essential to delivery – were holding up project implementation.

The adaptation and impact

In order to resolve this problem, a new mechanism was set up: a meeting every two weeks between the President, Chief of Staff, Minister of Finance and a few top advisers, to review progress on the President’s priority projects. This helped maintain momentum for the President’s top initiatives by ensuring that the key players in the system weren’t blocking progress. The high-level meeting complemented the activities of the core delivery unit, which continued to gather data and track progress on a wider set of projects.

Changing the System to Beat Ebola

Sometimes small changes are not enough. If the problem shifts sufficiently, the system itself needs to change too. Here we revisit Liberia’s response to Ebola.

The original mechanism

By the start of 2015, Liberia had made significant progress in tackling the epidemic. In early October 2014, there were an average of 60 new cases each day, and now Liberia was down to roughly a single new daily case, all in Greater Monrovia. This was in part thanks to the government’s coordination mechanism, the Incident Management System (IMS), which managed activities ranging from setting up treatment units to monitoring people who had been in contact with Ebola cases.

Why a change was made

It was becoming clear to Liberia’s crisis managers that this final stage of the epidemic was raising different challenges to those they faced at the peak of the epidemic several months earlier. First, all Ebola cases were now in Monrovia, an urban environment which posed particular challenges. Second, the nature of these later-stage cases was different and often more complex than earlier ones. And finally, the sheer number of organisations and people working on Ebola in the capital –international organisations, local non-governmental organisations, different levels of government – was itself a coordination challenge.

The adaptation and impact

The government, recognising that the existing IMS system was no longer fit for purpose, decentralised management of the response to four offices in different parts of Greater Monrovia. The idea was to bring management closer to the problem, and to try to coordinate better all the different organisations working on the response.

This decentralised approach emerged as an effective system for this stage of the epidemic, allowing for better engagement of communities that had been reluctant to work with government Ebola responders. By early May 2015, the World Health Organization declared Liberia Ebola-free, in part because of the work of this new coordination structure.



Sonpon Sieh appears to have been right not to take the advice of the advisers in that December 2014 meeting; their advice had little to do with the specific situation that Monrovia’s Ebola delivery system needed to manage.

In delivery we often get preoccupied with the form rather than the function of the system we’re building. But system building isn’t an exact science; the only way to get a delivery system that improves implementation is to start with the four building blocks of delivery, test different approaches, and then adapt them.

There’s room for improvement both on process – such as using a more structured PDIA approach, with learning mechanisms and feedback loops – and on documenting what types of delivery adaptations are worth testing in different contexts. We hope these examples from our work serve as a useful prompt for further discussion of these issues.

The work described here was carried out by the Tony Blair Africa Governance Initiative, it is now being continued by the Tony Blair Institute for Global Change.

Prioritisation in Government


Join us

Be the first to know what we’re doing – and how you can get more involved.